The present invention relates generally to systems and methods for providing secure and private communication channels for implantable medical devices using public communication networks and potentially insecure devices as intermediaries.
Implantable and programmable personal medical devices are becoming common. In some cases, these represent applications (“apps”) for smartphones. In other cases, regulated medical devices communicate through commercial off the shelf technologies. In both these cases, it is possible for provide communications to be intercepted, revealing potentially personal or confidential information. Further, it may be possible to modify communications, leading to misdiagnosis, improper treatment, or other harm.
For example, a smartphone app or the platform on which a smartphone app executes may include security vulnerabilities, even if the app is digitally signed. For example, of the medical device is a defibrillator, this could be reprogrammed to fail to generate a required discharge, or to discharge in a harmful or lethal pattern.
It is therefore desirable to make use of ubiquitous consumer technologies for communications, while isolating the privacy and security risks inherent in such technologies.
Øyvind Børthus and Tomas Mikael Engh, “Privacy protection in a mobile Biomedical Information Collection Service”, (Master's Thesis), Agder University College, Grimstad Norway (May 2005), expressly incorporated herein in its entirety, presents FIG. 1 which indicates a problem which persists to this day. How does one secure the communication from the patient to a provider, and from the provider to the patient, where the data communicated is private, and corruption of that data may be life-threatening? Børthus and Engh do not consider security of the “handheld device” as an issue to be addressed.
The use of a mobile wireless sensor gives the patient the possibility of being at home and doing his or hers normal daily activities while being monitored, but it also creates the need for new mechanisms for privacy protection. Fensli, R., E. Gunnarson, and T. Gundersen (2005) “A Wearable ECG-recording System for Continuous Arrhythmia Monitoring in a Wireless Tele-Home-Care Situation,” presented at The 18th IEEE International Symposium on Computer-Based Medical Systems, Dublin, Ireland. For the patient to stay in his normal environment gives several benefits for both the patient and the quality of the monitoring. The patient will be more relaxed and the recordings will not be affected by the stressful situation at a hospital.
A mobile wireless sensor is a compact electronic electrode attached to a patient that can measure different biomedical signals. In this scenario we are using a sensor for electrocardiogram (ECG) recordings as an example. This sensor will continuously measure and wirelessly transmit sampled ECG-recordings using a built-in RF-radio transmitter. The RF-radio receiver converts the ECG-samples by the use of a microcontroller before transmitting the ECG-samples to a standard personal digital assistant (PDA). The sensor measures ECG-signals with a sampling frequency up to 1000 samples per second. Crawford, M. H. et al. (1999) “ACC/AHA Guidelines for Ambulatory Electrocardiography: Executive summary and Recommendations” American College of Cardiology. The signal is digitalized with 10 bit resolution, requiring up to 10 kb/s of bandwidth plus overhead to transmit to the PDA. The range of the RF-radio signal is up to 10 meters. The transmitter chip used by the sensor described in Fensli et al. is a RF-transmitter CC1050 from ChipCon, operating at 869.700 MHz, with a bit rate of up to 76.8 kbit/s. The sensor will transmit continuously, and will be attached to the patient for 3 days to a week at a time. The sensor is a disposable unit, and will only be used once.
A PDA is used to receive the information from the sensor of Fensli et al. and will often have 400 to 600 MHz processor, 64 to 128 MB of internal memory and a memory card with capacity of more than 2 GB for storing data. The PDA has both a short range wireless RF-radio device and a GPRS card installed. The PDA is an “intelligent” unit, using automatic arrhythmia detection algorithms for analyzing the signals from the sensor and decides if the recordings are within normal values. As long as the signals are within normal values the PDA will regularly send an extract of the recordings to an electronic health register (EHR) connected to the Internet by the use of GPRS communication. If an abnormal ECG activity is encountered, the PDA will store 1 minute of the ECG recordings and then transmit the recordings to the EHR server.
There are several security and privacy threats relevant to this scenario. From a medical standpoint it is very important that the PDA only receives data from the correct sensor, and not from other devices in the same area. Keeping the integrity of the data is also very important to avoid false alarms or incorrect normal signals. The data must also be protected in such a way that possible attackers cannot get access to personal information about the patient.
For the transmission of data from the PDA to the EHR, and between users in the national health network many of the same threats applies. Integrity, security and privacy must be protected when sending electronic messages containing sensitive medical and personal information. The electronic messages contain a lot more sensitive information than the data packets sent from the sensor and thus require a higher level of security protection. Non repudiation is especially important when sending messages regarding medication and diagnosis.
Bluetooth is a short range radio standard designed for low power consumption. On Jun. 14, 2002 the Bluetooth standard was published as IEE 802.15.1. Subsequent versions have been published. Bluetooth operates in the unlicensed industrial, scientific and medical (ISM) band at 2.4 to 2.485 GHz, using a spread spectrum, frequency hopping, full-duplex signal at a nominal rate of 1600 hops/sec. Bluetooth SIG (2004) “Specification of the Bluetooth System”. The adaptive frequency hopping (AFH) was designed to reduce interference between wireless technologies using the 2.4 GHz spectrum. The AFH detects other devices and avoid the frequencies used and can take advantage of the available frequencies. The AFH hops among 79 frequencies at 1 MHz intervals to give high interference immunity. The range of Bluetooth devices range from 1 to 100 meters depending on the device class used. Class 1 devices use a power of 100 mW giving it a range of up to 100 meters, class 2 uses 2.5 mW and have a range of up to 10 meters, and class 3 uses 1 mW and have a range of up to 1 meter. Class 2 is the most commonly used. Version 1.2 has a gross data rate of up to 1 Mbit/s and Version 2.0 with Enhanced Data Rate (EDR) has a gross data rate of up to 3 Mbps. The effective transfer rates are respectively 723.1 kbit/s and 2.1 Mbit/s. Bluetooth devices communicate in small groups of up to 8 devices, called a piconet. One device plays the “master” and the rest (up to 7) devices are “slaves”. Data can be transferred between master and 1 slave at any given time, and the master switches between slaves in a round robin fashion. Simultaneous transmission from master to multiple slaves is possible, but rarely used.
A Bluetooth device will transmit the following sets of information on demand: Device name; Device class; List of services; and Technical information. Any device can perform an inquiry to find other devices with which to connect, and any device can be configured to respond to such inquiries. If the device trying to connect knows the address to the device, it will always respond to inquiries with the information mentioned above if requested for it. Use of the devices services may require pairing or owner accept, but the connection itself can be established by any device and held until it goes out of range. All devices have a unique 48 bit address, but these are generally not shown in inquiries and instead user friendly “Bluetooth names” are used. These names can be set by the user and most devices come with a standard name set by the manufacturer. All devices also have a 24 bit class identifier, providing information about what class of device it is, for example mobile phone, headset or computer.
Bluetooth devices can be paired to establish a trusted connection. By user input (a pin code) they can learn a shared secret key known as a “passkey”. A device can then cryptographically authenticate the identity of another device. With some devices, like wireless earphones, it is impossible for the user to enter a pin code, and the device has a fixed pin code, which can be entered into the peer device. Trusted devices can also encrypt information they transmit so no one can “listen in”. The encryption can be turned off, and the passkey is stored in the device's memory, and not in the Bluetooth chip itself. The trusted connection can be canceled by either device at any time. Devices will generally require pairing or user input before it allows a remote device to use its services.
In order to communicate with other Bluetooth devices a device must be able to interpret certain Bluetooth profiles. These profiles define the possible applications. 24 profiles are defined and adopted by the Bluetooth SIG, for example: Generic Access Profile (GAP) which provides the basis for all other profiles. This profile describes which features must be implemented in all Bluetooth devices, generic procedures for discovering and linking to devices, and basic user-interface terminology; Basic Imaging Profile (BIP). This profile is designed for sending images between devices and includes the ability to resize, and convert images to make them suitable for the receiving device; and Human Interface Device Profile (HID) provides support for devices such as mice, joysticks, keyboards, etc.; Advanced Audio Distribution Profile (A2DP). Also referred to as the AV profile, it is designed to transfer a stereo audio stream like music from an MP3 player to a headset or car radio.
On the link layer Bluetooth uses the SAFER+ algorithm for authentication and key generation, and E0 stream cipher for encrypting packets. The link layer security is independent of possible application layer security. The SAFER+ (Secure and Fast Encryption Routine) algorithm is a block cipher with block size of 128 bits, and a default key size of 128 bits. Wikipedia.org “SAFER” en.wikipedia.org/wiki/SAFER. The cipher uses 8 rounds with 4 stages; a key-mixing stage, a substitution layer, another key-mixing stage, and finally a diffusion layer. The E0 is a stream cipher. Wikipedia.org “E0 (cipher)” en.wikipedia.org/wiki/E0_(cipher). It generates a sequence of pseudorandom numbers and combines it with the data using a XOR operator. The key length is usually 128 bits, but may vary. For each iteration E0 generates a bit using 4 shift registers of different length (25, 31, 33, 39 bits), and two internal states, each 2 bits long. For each clock cycle, the registers are shifted and the two states are updated with the current state, the previous state and the values in the shift registers. Four bits are then extracted from the shift registers and added together. Then the algorithm XORs that sum with the value in the 2-bit register. The first bit of the result is output for the encoding. E0 is divided in three parts: 1. Payload key generation; 2. Key stream generation; and 3. Encoding.
The setup of the initial state in Bluetooth uses the same structure as the random bit stream generator. We are thus dealing with two combined E0 algorithms. Using the 128-bit key, Bluetooth address on 48 bits and the 26-bit master counter an initial 132-bit state is produced at the first stage. The output is then processed by a polynomial operation and the resulting key goes through the second stage, which generates the stream used for encoding. The key is a multiple of 2 varying from 8 to 128 bits length. 128 bit keys are generally used. These are stored into the second stage's shift registers. 200 pseudorandom bits are then produced, and the last 128 bits are inserted into the shift registers. It is the stream generator's initial state.
Shaked and Wool showed (Shaked, Y. and A. Wool (2005) “Cracking the Bluetooth PIN” www.eng.tau.ac.il/˜yash/shaked-wool-mobisys05/index.html (current May 2, 2005)) that the PIN code used by some devices to add security can be easily broken, even on an old computer. They described a passive attack, in which an attacker can find the PIN used during the pairing process. They used less than 0.3 seconds to crack a 4 digit pin, as used by most devices using a pin code, on a Pentium III 450 MHz computer and even faster on a new 3 GHz Pentium IV. If two Bluetooth devices perform pairing in a hostile area, they are vulnerable to this attack. Lu, Meier, and Vaudenay (Lu Y., W. Meier, and S. Vaudenay (2005) “The Conditional Correlation Attack: A Practical Attack on Bluetooth Encryption” Crypto '05, Santa Barbara. [19] CABA “Standards and protocols” www.caba.org/standard/zigbee.html (current Jan. 18, 2006)) presented an attack on the E0 stream cipher. Using a conditional correlation attack developed and optimized against Bluetooth two-level E0 they attack a recently detected flaw in the resynchronization of E0. Their best attack finds the original encryption key for two-level E0 using the first 24 bits of 223.8 frames and with 238 computations. This is the fastest and so far only practical known-plaintext attack on Bluetooth encryption.
ZigBee Alliance was incorporated in August 2002, and announced the ZigBee standard in December 2004. The ZigBee standard was released public in June 2005. Other wireless transfer methods are focusing on transferring large amount of data as fast as possible, but ZigBee is going the other direction. It focuses on low powered devices with a need for security and sending small amounts of data. The ZigBee standard was created to “enable reliable, cost-effective, low-power, wirelessly networked, monitoring and control products based on an open global standard” ZigBee Alliance “Our Mission”. www.zigbee.org/en/about/ (online May 25, 2006). The most common devices that use ZigBee are industrial automation, remote metering, embedded sensors, medical devices, smoke and intruder alarms, interactive toys, building automation and home automation. ZigBee operates in the European 868 MHz ISM band with one channel, the American and Australian 915 MHz ISM band with 10 channels or the 2.4 GHz ISM band with 16 channels. The data rate is 250 kbit/s in the 2.4 GHz band, 40 kbit/s in the 915 MHz band, and 20 kbit/s in the 868 MHz band. ZigBee Alliance (2004) “ZigBee Specification v1.0”; Kinney, P. (2003) “ZigBee Technology: Wireless Control that Simply Works”, Kinney Consulting LLC, Chair of IEEE 802.15.4 Task Group, Secretary of ZigBee BoD, Chair of ZigBee Building Automation Profile WG. Transmission range is typical between 10 and 75 meters. The ZigBee protocol supports up to 65,536 nodes. It has handshaking for transfer reliability.
The ZigBee stack architecture is based on the standard seven layer c is controlling the access to the radio channel. The mechanism used is called CSMA-CA. The MAC layer also transmits beacon frames, synchronization, and provides reliable transmission mechanisms. The network layer has 3 main functions: join and leave networks, apply security, and route frames to their destinations. In a coordinator device, the network layer has the responsibility to start a new network and discover what kind of application services nearby devices. It also assigns addresses to newly assigned devices. The network layer supports star, cluster three and mesh topology. As mentioned above, the APL layer consists of Application Support sub-layer (APS), ZigBee Device Object (ZDO) and manufacturer-defined applications. The APS is responsible for maintaining the tables for binding i.e. the ability to match two devices and forward the messages between two devices. The ZDO define the role of a device in the network (network coordinator, coordinator, or end device), initiate and/or respond to binding requests, and establish a secure connection.
ZigBee has several different security mechanisms (ZigBee Alliance (2005) “ZigBee Security specification overview”), and are found in the MAC layer, NWK layer and the APS layer. Among them are freshness, integrity, authentication, and encryption. The freshness checks prevent replay attacks. It uses incoming and outgoing freshness counters that are reset every time a new key is created. The integrity checks prevent anyone from modifying the message, and supports up to 128 bit message integrity. Authentication is handled either in the network level or the device level. The network level authentication is achieved when using a common network key. This will prevent attacks from outsiders, and it has very little memory cost. The device level authentication is achieved when using unique link keys between pair of devices. This prevents attacks from both outsiders and insiders, but has a higher memory cost. ZigBee supports 128 bit AES encryption. This encryption can be used either at network level or device level, and is handled the same way as authentication. The encryption can be turned on or off without impacting the freshness, integrity or authentication. ZigBee can also add security to frames. ZigBee Security can add headers to the data frames at the MAC, NWK, and APS layers. ZigBee supports 3 types of topologies: Star, Cluster tree and Mesh.
In a star topology the network is controlled by a PAN coordinator (network controller). All end devices can only talk to the coordinator. The coordinator is almost always in a listening mode, except when new end devices are trying to connect. The star topology supports up to 65,536 end devices. It is a very simple layout and has low latency. Øyen, G. E. (2006) “ZigBee and IEEE 802.15.4: A brief introduction”. In a cluster tree topology the tree structure is rooted at the PAN coordinator. The coordinator initiates the network, and the children (end devices) routes through parents in a hierarchy. It uses a multi-hop topology to increase the network range. The cluster tree topology is not ideal for network devices that require low latency. The idea with mesh topology is that messages can be routed from any source to any destination. The way this is done is that every FFD is functioning as a router for all its neighbors. Like cluster tree topology, the mesh topology uses multi-hop topology to increase the network range. It has high reliability, since the messages can go many routes. If one or more of FFD disconnects, the messages still gets to the destination, but uses another route than it normally does. This way it is self-configuring. Since this topology depends on the routers, it may not be ideal for battery driven networks, as the routers will have relatively large power consumption.
Only one Zigbee coordinator (ZC) is required in each ZigBee network. It is the most capable device in the network, and initiates the formation. It is the root of the network tree, and might bridge to other networks. It acts as a PAN coordinator (FFD) and as a router when the network is formed. The ZC also acts as a repository for security keys. The coordinator is also assumed to be the trust center, which is responsible for allowing new devices into the network and for distributing keys. It is possible for the trust center to be a dedicated device. The Zigbee Router (ZR) is an optional component in a ZigBee network. The routers associate with the ZC or with other previously associated ZR. The ZR acts as a coordinator (FFD) and is used as a local address allocation/de-allocation device. It is used in multi-hop routing of messages. The ZR also looks after its own Zigbee End Device (ZED). The ZED contains very little functionality. It is limited to communicate with its coordinator. The ZED is not allowed to associate or participate in routing. It requires the least amount of memory and is therefore cheaper than ZC or ZR. It has low power consumption since its parent puts it to sleep.
When a new device is installed in the network, it will initiate queries to discover already active ZigBee devices in the network. The request is either an IEEE address request, which is unicast, or a NWK address request, which is broadcast. When the unicast request is sent, it assumes the NWK address is known. When the broadcast request is sent, it carries the known IEEE address as payload. The response on these queries is dependant on the three device types mentioned above: ZED, ZR and ZC. The ZED responds to the query by sending its own IEEE or NWK address. The ZR responds to the query by sending its own IEEE or NWK as well as the IEEE and NWK address of all the other devices connected to the ZR. The ZC responds to the query by sending its own IEEE or NWK as well as the IEEE and NWK address of all the other devices connected to the ZC.
There are 3 different key types used in ZigBee; master key, link key and network key. The master key is used as basis for long term communication between devices, and can be either factory installed or be set up over the air or using out-of-band mechanisms. The link key is used for security between two devices. The link key is also used to authenticate devices to the coordinator device. The network key is used for security in a network. The link and network keys can be factory installed, be set up using a symmetric key-key exchange handshake or be sent from the trust center.
A wide range of ZigBee transceivers that are suited for use in a wireless sensor are available on the commercial market, from suppliers like ChipCon, CompXs, Helicomm Inc. and others. Newer models can come with built in hardware support for data encryption and authentication using AES on the link layer. An example of such a chip is the ChipCon CC2510. ChipCon “CC2510 Product Information” www.chipcon.com/index.cfm?kat_id=2&subkat_id=12&dok_id=258 (online May 24, 2006). The CC2510 is a powerful 2.4 GHz ISM band System-on-Chip designed for low-power and low-voltage wireless communication applications. This chip includes a dedicated 128-bit AES coprocessor to minimize the MCU usage when encrypting. It also has a dedicated DMA controller which moves data from a peripheral (in our case the sensor) to the memory with almost no intervention from the MCU. This way the MCU workload is reduced to a minimum. The chip has a 32 KB of programmable flash memory and 4 KB of RAM. This chip is developed to be energy efficient and have a low unit cost.
The ZigBee device installed in the PDA will most likely be an SD or CF card. Since our PDA must have a GPRS card, which usually is a CF card, the ZigBee card will be an SD card. An example of a ZigBee SD card is produced by C-Guys. C-GUYS www.c-guysusa.com (online May 24, 2006). They are a company that specializes in developing different SD and SDIO devices, such as SD controllers, memory cards, adapters and card readers. One of their products is a ZigBee SDIO card for use in PDA. The SDIO card is using the standard ZigBee frequency, the 2.4 GHz ISM band, has 250 kbps data rate and 10 meter range.
GSM offers several security services (Schiller, J. (2003) Mobile Communications, second edition, Addison-Wesley, pages 93-156), and they are found either in the SIM card or the AuC (authentication center, a separate system in the network that contains the algorithms for authentication and the keys for encryption). The SIM card stores personal data and a secret key Ki, and is only accessed with a four digit PIN number. After MS authenticates itself, the MS and BTS (base transceiver station) encrypts all voice and data. There are 3 types of algorithms: A3 for authentication, A5 for encryption and A8 for generation of the cipher key. The algorithms are very weak, but it is possible for the network providers to use stronger algorithms for encryption or user can provide stronger end-to-end encryption. To encrypt the messages, a key Kc is created by using the individual key Ki and a random number by generated by the A8 algorithm. The Kc key is calculated both in the MS (SIM) and the network, and is not transmitted over the air interface.
GPRS (general packet radio service) is a packet-oriented operation in the GSM system, often called 2.5G, since the technology lies somewhere between 2G and 3G (2nd and 3rd generation mobile technology). en.wikipedia.org/wiki/GPRS (current May 24, 2006). The idea behind GPRS is Wikipedia.org “GPRS” that all or some GSM-channels (time slots) are combined to one channel with higher capacity. Schiller, J. (2003) Mobile Communications, second edition, Addison-Wesley, pages 93-156. While GSM was primarily designed for voice transmission, GPRS is a more data-oriented transmission. The transmission is packet-oriented so that many users can transfer data when they need it, and don't use bandwidth when they have nothing to send. This type of transmission is especially designed for frequent transmission of small volumes of data, e.g., typical web requests or web response. The overall goal is the provision of a more effective and cheaper packet transfer for typical Internet applications that rely solely on packet transfer. The ISP is usually taking charge for the data volume transferred instead of charging for the connection time. Andersen P. B. og R. Johnsen “Mobiltelefon—Ikke bare prat” in (Ed.) Kunnskapsforlagets Årbok 2000 fag.grm.hia.no/ragnarj/mobile_syst/tradlos_komm.pdf. By doing it this way, the user is “always on”, and no connection has to be set up for the transfer.
For each new GPRS radio channel, the GSM can locate between one and eight time slots within a TDMA frame. Schiller, J. (2003) Mobile Communications, second edition, Addison-Wesley, pages 93-156. All time slots can be shared by the active users. It is possible to get a transfer rate to 170 kbit/s if you have max slots and are using the fastest coding, but a more realistic bit rate is 30-80 kbit/s. Wikipedia.org “GPRS” en.wikipedia.org/wiki/GPRS (current May 24, 2006). CS-4 is the fastest coding scheme, but the least robust. This coding is available near the Base Transceiver Station (BTS). CS-1 is the slowest coding scheme but is most robust and is available when the Mobile Station (MS) is further away from the BTS. The GPRS operators usually reserves at least one time slot per cell to guarantee a minimum data rate. Users can specify a QoS-profile themselves. This determines the service precedence (high, normal, low), reliability class and delay class of the transmission and user data throughput. The latency of GPRS is incurred by channel access delay, coding for error correction, and transfer delay in the fixed and wireless parts of the network. Due to these parts involved, the latency is very high, even with small packet sizes. A round trip is often higher than 1 sec, even with packets as small as 128 byte. Table 2 shows some examples of the delay classes with different service data units (SDU) sizes. A MS (Mobile Station) that are using GPRS are considered a part of the internet and are assigned a private IP address. The operator translates the IP address into global addresses at the GGSN (Gateway GPRS support node) using a NAT (Network Address translation). The advantage of this approach instead of giving the MS an “ordinary” IP address is to protect the MS against attacks. The private IP addresses are not routed through the internet, so it is impossible to reach an MS from the internet. Other security mechanisms are the same as GSM.
EDGE (enhanced data rates for GSM evolution) is a digital enhancement of GSM, and the next step towards 3G and UMTS (EDGE is also called 2.75G, Wikipedia.org “2.75G” en.wikipedia.org/wiki/2.75G (current Mar. 23, 2006)). EDGE is using an enhanced modulation scheme and other techniques to get data rates up to 384 kbit/s using the same carrier size and frequencies as GSM. EDGE does not require any changes in the GSM core networks, but the base stations have to be upgraded. Wikipedia.org “EDGE” en.wikipedia.org/wiki/EDGE (current May 24, 2006). Besides better data rate, the most important addition to GSM is CAMEL (customized application for mobile enhanced logic). CAMEL is an intelligent network support, and the services offered are especially effective when a subscriber is roaming between international network operators. Meskauskas, P. “Customised Applications for Mobile Enhanced Logic (CAMEL)” Examples are no-prefix dialing and seamless MMS messages from other countries. EDGE is backward compatible with GPRS, and will use GPRS as transmission in those areas without EDGE support. Norway's largest telecommunication company, Telenor, has very good national EDGE coverage (Telenor Dekningskart telenormobil.no/dekninginnland/index.do (online May 25, 2006) www.telenor.no/bedrift/produkter/mobil/merom_umts_edge.html), and offers a download rate of 100-200 kbit/s and upload rate of 50-75 kbit/s.
UMTS (universal mobile telecommunication system) is a third generation mobile technology used in Europe and Japan, and the 3G successor of GSM. Wikipedia.org “UMTS” en.wikipedia.org/wiki/UMTS (current May 23, 2006). To avoid very high implementation cost, UMTS try to reuse as much GSM/GPRS technology and infrastructure as possible. It is especially designed for high-speed services like video telephony. All signals use the same frequency band, a 5 MHz wide band licensed to network operators. The signals are multiplied with a chipping sequence which is unique to each user. If someone tries to tap the signal, it would appear as noise to him if he don't know the spreading code. In its initiation phase, UMTS has a theoretical bit rate of 384 kbit/s in high mobility situations, and up to 2 Mbit/s in stationary user environments. UMTS Forum “What is UMTS?” www.umts-forum.org/servlet/dycon/ztumts/umts/Live/en/umts/What+is+UMTS_index (online May 25, 2006). It takes twice as many base stations as GSM to achieve the same coverage, and to get fully fledged UMTS features including Video on Demand, a base station need to be set up every 1-1.5 km. Some of the downsides of UMTS (as it is today) are: very poor coverage, poor battery life on the MS, impossible to provide complete UMTS features in rural areas, and lack of consumer demand for 3G.
Hash functions are the most versatile of all cryptographic primitives. Bishop, M. (2003) Computer Security Art and Science, Addison-Wesley. It can be used for encryption, authentication, and a simple digital signature. The typical use of a hash function is digital signatures. The idea behind hashing is to take a long string of bits (or bytes) as input, run a hash function, and produce a fixed length hash sum. Mao, W. (2004) Modern Cryptography Theory & Practice, Bristol, Prentice Hall. If you have a message (m) and a hash (h), you are signing h(m) instead of signing m. The reason for signing h(m) is that the message (m) are usually very large, up to millions of bits, but the hash function are usually between 128 and 256 bits, thus making it much faster and more effective. One of the practical problems with selecting a hash function, is that there's only a couple methods to choose from Wikipedia.org “Cryptographic Hash Function” en.wikipedia.org/wiki/Cryptographic_hash_function (current Apr. 25, 2006); the SHA family and MD5. There are a couple of alternatives, but they have not been tested thoroughly enough to trust them. A typical hash function is shown below. MD5 is a cryptographic hash function used to verify data integrity. Rivest, R. (1992), MIT laboratory for Computer Science and RSA Data security, inc. April 1992. It was developed by Ronald Rivest in 1991 to replace MD4, because MD4 proved to have some security weakness. When using MD5, the message is split into blocks of 512 bits. Answers.com “MD5” www.answers.com/topic/md5#after_ad1 (online May 25, 2006). The last block is padded, and includes the length of the message. MD5 has 128-bit hash value that is split into four words of 32 bits, each with a compression function h′ with four rounds. Each round mix the message block and the state, with a combination of addition, XOR, AND, OR and rotation operations on 32-bit words. This way each message word is used four times. After the four rounds of the h′ function, the input state and the result are added together to produce the output of h′. The structure of operating 32-bits words is very efficient on 32-bits CPUs.
One of the basic ideas behind hash functions is that it is collision resistant. As of 2006, there were no known attacks on the MD5 function, but a collision of the compression function occurred in 1996. For modern applications, the 128-bit hash size is insufficient, and it is possible to find real collisions in about 264 evaluations of the hash function. This made security experts to recommend a replacement. One of them was SHA-1. SHA (Secure Hash Algorithm, Wikipedia.org “SHA Hash functions” en.wikipedia.org/wiki/SHA (current May 6, 2006)) is a set of cryptographic hash functions. The first standard was just called SHA, but is now referred to as SHA-0. It was developed in 1993 by NSA (National Secure Agency) and published as a US government standard. It was found a weakness in this function, and NSA developed a fix which they published as an improved version called SHA-1.
SHA-1 (National Institute of Standards and Technology (1995) “Secure Hash Standard” www.itl.nist.gov/fipspubs/fip180-1.htm (online May 25, 2006)) is the successor of SHA-0 (and MD5) (Wikipedia.org “SHA Hash functions” en.wikipedia.org/wiki/SHA (current May 6, 2006)), and was developed in 1995. It is used in a wide area of security applications, like TLS, SSL, PGP, SSH, S/MIME, and IPSec. SHA-0 and SHA-1 is based on the same principles as MD4 and MD5, and produces a 160 bit message digest with a maximum size of 264 bits. It is, unfortunately, 2-3 times slower than MD5. SHA-1 has a 160-bit state consisting of five 32-bit words. It uses four rounds that consist of a mixture of 32-bit operations. SHA-1 uses a linear recurrence to stretch the 16 words of a message block to the 80 words it needs, to ensure that each message bit affects the mixing function at least a dozen times.
The main problem with SHA-1 is the 160-bit result size. Collisions can be generated in only 280 steps, but it is reported that it can be generated in as few as 263 steps. Schneier, B. (2005) “New Cryptanalytic Results Against SHA-1” www.schneier.com/blog/archives/2005/08/new_cryptanalyt.html (current Aug. 17, 2005).
3TDES has following specification: Three 56-bit DES keys=168 bits+parity bits=192 bits. The effectiveness is counted as only 112 bits because of the exposure to the man-in-the-middle attacks. The best know attack on the 3TDES requires 232 known plaintexts, 2113 steps, 290 single DES encryptions and 288 bit memory. 3TDES is no longer considered a very good encryption method. It is being replaced by its successor, AES with its better security mechanisms. AES has a larger block size, longer keys, freedom from crypto analytic attacks, and proves to be up to six times faster than 3TDES.
Advanced Encryption Standard (AES, Wikipedia.org “Advanced Encryption Standard” en.wikipedia.org/wiki/Advanced_Encryption_Standard (current May 24, 2006)), also known as Rijndael is a block cipher. In 2000 NIST, National Institute of Standards and Technology, chose Rijndael as the new encryption standard for the US government. NIST selected the Rijndael algorithm in front of 4 other competitors based on the combination of security, performance, efficiency, ease of implementation and flexibility. c Rijndael is a block cipher and uses a substitution-linear transformation network with 10, 12 or 14 rounds, depending on the key and block size. The key and block size can be individually specified to 128, 192 or 256 bits. A data block to be encrypted by Rijndael is split into an array of bytes, and each encryption operation is byte-oriented. Only block size of 128 bits is adopted in the AES standard. The AES does not describe how to handle and distribute keys, and need a secure key management infrastructure to maintain its high level of security. AES is about 6 times faster than 3DES in software implementations.
AES is a block cipher with a fixed block size of 128 bits and a variable key size of 128, 192 or 256 bits. The 128 bit message input block is segmented into 16 bytes. The data structure for their internal representation is a 4×4 matrix. Like the DES algorithm, AES comprises a plural number of iterations of a basic unit of transformation: “round”. Depending on the size of the key, AES uses 10, 12 or 14 rounds. A round transformation in AES is denoted by: Round (State, RoundKey). State is a round-message matrix, and is treated as both input and output and a length of 128 bits; RoundKey is a round-key and is derived from the input-key via key schedule. All round-keys are 128 bits, also when the encryption key is 192 or 256 bits. The key schedule is an algorithm for computing the sub-keys for each round in a product cipher from the encryption (or decryption) key. The execution of a round will cause the elements in of State to change value. For encryption the State input to the first round is the plaintext message matrix, and the output from the last round is the cipher text message matrix. For decryption they are respectively cipher text and plaintext message matrix. Each round, except the last round, consists of 4 stages: 1. SubBytes; 2. ShiftRows; 3. MixColumns; and 4. AddRoundKey. All rounds are identical with the exception of the final round, which does not include the MixColumns transformation. The round transformations are invertible for the purpose of decryption.
The SubBytes function provides a non-linear substitution on each byte of State. In the SubBytes step, each byte in the array is updated using an 8-bit S-box. A substitution box (or S-box) is a basic component of symmetric key algorithms, and takes some number of input bits, m, and transforms them into some number of output bits, n. This operation provides the non-linearity in the cipher. The S-box used is derived from the inverse function over GF(28), known to have good non-linearity properties. Non-linearity is an important property for a block cipher to prevent differential cryptanalysis. The ShiftRows step operates on the rows of the state; it cyclically shifts the bytes in each row by a certain offset. For AES, the first row is left unchanged. Each byte of the second row is shifted one to the left. Similarly, the third and fourth rows are shifted by offsets of two and three respectively. In this way, each column of the output state of the ShiftRows step is composed of bytes from each column of the input state. In the MixColumns step, the four bytes of each column of the state are combined using an invertible linear transformation. The MixColumns function takes four bytes as input and outputs four bytes, where each input byte affects all four output bytes. Each column is treated as a polynomial over GF(28) and is then multiplied modulo x4+1 with a fixed polynomial c(x)=3x3+x2+x+2. ShiftRows and MixColumns are intended to achieve a mixture of the bytes positioned in different places of a plaintext message block. In the AddRoundKey step, the subkey is combined with the state. For each round, a subkey is derived from the main key using the key schedule; each subkey is the same size as the state. The subkey is added by combining each byte of the state with the corresponding byte of the subkey using bitwise XOR. This stage provides necessary secret randomness to the message distribution. To decrypt, the 4 functions of each round are inverted, and implemented in reverse order. The AddRoundKey is its own inverse and the same for both encryption and decryption. Three different key lengths, 128,192, and 256 are supported by AES, making it possible to choose stronger security or better efficiency. All key lengths are secure enough to be used for most levels of classified information. Only for extreme security requirements is it required to use 192 or 256 bit key lengths. The American National Security Agency has conducted a research on the strength of the AES algorithm (CNSS (2003) “CNSS Policy No. 15, Fact Sheet No. 1 National Policy on the Use of the Advanced Encryption Standard (AES) to Protect National Security Systems and National Security Information”) stating: “The design and strength of all key lengths of the AES algorithm (i.e., 128, 192 and 256) are sufficient to protect classified information up to the SECRET level. TOP SECRET information will require use of either the 192 or 256 key lengths.”
Public Key Infrastructure (PKI) is a policy to establish a secure method for information exchange. Bishop, M. (2003) Computer Security Art and Science, Addison-Wesley. It is also a set of integrated services and administrative tools to create and manage applications based on public keys. This includes cryptographic methods, the use of digital certificates, certification authorities, and the system to manage the process. There are two key elements in PKI: Public Key Cryptography and Certification Authorities (CA). Public Key Cryptography is a form of cryptography and uses a pair of cryptographic keys designed as a private key and a public key, which are related mathematically. The private key is kept secret by the user and the public key may be widely distributed. Generally, if user Bob shall send a message to user Alice, Bob will contact Alice and ask for her public key. Alice sends Bob her public key, and Bob uses it to encrypt his message. Bob will then send Alice the message, encrypted with Alice's public key, and the only way to decrypt the message is to use Alice's private key. Børthus, B and Tomas, E. (2005) “Public Key Infrastructure for Windows Server 2003”.
Some examples of public key techniques are: Diffie-Hellmann, DSS, ElGamal, RSA, and various Elliptic Curve techniques. Wikipedia.org “Public Key Infrastructure” (current May 26, 2006).
A CA is responsible for establishing and vouching for the identity of certificate holders. A CA also revokes certificates if they are no longer valid and publishes certificate revocation lists (CRLs) to be used by certification verifiers. The certificates are issued by a CA based on information provided in the certification request and settings contained in a certification template. A certification template is the set of rules and settings that are applied against incoming certificate requests. The most common digital certificates in PKI use the X.509 Digital Certificate format and usually contain the following: The user's public key; The user's identity, such as name and e-mail address; The validity period of the certificate; The digital signature of the issuer, which attest to the validity of the binding between the user's public key and the user's identifier information. There are different levels of certificates based on the need for functions. As a general rule, the higher level of the certificate, the stricter are the policies for verifying. PKI supports hashing to keep the integrity of the data.
Smart cards are pocket sized plastic cards with embedded integrated circuits. There are 2 broad categories of cards; memory cards and microprocessor cards. c. The standardization of smart card systems is an ongoing process. One of the standards most referred to is the ISO-7816 standard. A memory card contains non-volatile memory that can store information and perhaps some specific non-programmable security logic. An example of a memory card is a prepaid phone card. They can also be used as a high security alternative to magnetic stripe cards. Memory cards can only perform fixed operations. Microprocessor cards contain memory and microprocessor components. These cards can process data on the card and can used for a variety of applications. Microprocessor cards can provide secure access to networks, be used as SIM card in mobile phones and as electronic wallets. Smart cards are engineered to be tamper resistant and are very suitable to hold personal digital signatures that can be used as authentication to grant access to secure networks. Hong Kong University of Science & Technology (1998) “Guide to Smart Card Technology”.
The definition of VPN is “A virtual private network (VPN) is a private data network that makes use of the public telecommunication infrastructure, maintaining privacy through the use of a tunneling protocol and security procedures”. VPN Consortium (2006) “VPN Technologies: Definitions and Requirements” www.vpnc.org/vpn-technologies.html. A VPN makes it possible to share resources in a secure way over an insecure public network like the Internet. There are 3 important VPN technologies used: secure, trusted and hybrid VPN. Only secure VPN is relevant to our scenario. Secure VPN uses an encrypted secure “tunnel” to transport data over a public network. Tunneling is generally done by encapsulating the private network data and protocol information within the public network transmission units so that the private network protocol information appears to the public network as data. Tunneling allows the use of the Internet, to convey data on behalf of a private network in a secure way. There are several secure VPN protocols, like IPsec, SSL and PPTP. A properly chosen, implemented, and used secure VPN protocol can provide secure communications over unsecured networks, and provide protection of confidentiality and integrity, and sender authentication to ensure privacy.
Secure authentication is very important when using a VPN solution. Authentication mechanisms can make use of what you know (pin code, password), what you have (smart card) or what you are (fingerprint, retinal scan). Wikipedia.org “Virtual private network” en.wikipedia.org/wiki/Virtual_Private_Network (current May 25, 2006). The use of one of the above will give weak authentication, but the use of two will give a much stronger authentication.